Roxadustat: A catalyst for diabetic wound healing through re-epithelialization and angiogenesis

Ethics statement

This animal experiment followed the requirements of the ethical review system for the experimental animal welfare of Wuhan Myhalic Biotechnology Co., Ltd. Animal Ethics Committee. Approval number HLK-20210906-001. The mice were fed under specific pathogen-free conditions, and they underwent euthanasia after sampling.

Induction of diabetic mice

Male BALB/c mice aged 10–12 weeks (20 ± 2 g) were purchased from Beijing Weitong Lihua Laboratory Animal Technology Co., Ltd (Beijing, China). The mice were certified to be of specific-pathogen-free grade (Animal Quality Certificate Number: No. 110011241101163626).

The mice received a single intraperitoneal injection of streptozotocin (STZ; 160 mg/kg; 18883-66-4, Sigma, USA).^[17] After 72 h, their blood glucose level was tested through the tail pricking method, and mice with blood glucose higher than 16.7 mmol/L were recognized as diabetic mice. The diabetes mice were fed for 4 weeks before wound modeling. The healthy mice from the same batch served as a control group.

Full-thickness skin excisional wound model

After anesthesia by intraperitoneal injection of 1% pentobarbital (P5178, Sigma, USA) at 40 mg/kg and hair removal were performed, two full-thickness skin wounds of 4 mm in diameter were made on the back of each mouse as described in the previous study.^[14] A self-adhesive silicone ring was placed around the wound to offset wound shrinkage. This silicone inner ring can be a reference for measuring the wound size. Then, all mice were housed individually.

Animal experiment grouping scheme and treatment measures

In accordance with previous studies,^[15] roxadustat (S1007, Selleck, USA) was first dissolved in dimethyl sulfoxide at a concentration of 50 mg/ml and further diluted to 1 mg/mL in sterile phosphate-buffered saline (PBS). The mice were randomly divided into three groups, and they received corresponding treatments.

Diabetes + FG-4592 group (n = 6; 12 wounds in total): This group was composed of STZ-induced diabetic mice fed for 4 weeks after successful induction. Roxadustat was injected intraperitoneally (IP) daily at a dose of 10 mg/kg after the wound model was taken.

Diabetes group (n = 6; 12 wounds in total): This group consisted of STZ-induced diabetic mice fed for 4 weeks after successful induction. An equal amount of PBS was injected IP daily after a wound model was established.

Control group (n = 6; 12 wounds in total): This group comprised normal mice from the same batch as the other two groups. An equal amount of PBS was injected IP daily after a wound model was established.

On days 0, 4, 8, and 12 after injury, blood glucose was monitored through the tail vein, and mouse body weight was measured with a balance. Half of the mice (n = 3) in each group were sampled and sacrificed on day 6 after injury; another half (n = 3) was sampled and sacrificed on day 14 after injury. At the end of the experiment, the mice were euthanized by decapitation after pentobarbital anesthesia. The pentobarbital was administered IP at a concentration of 50 mg/mL and a dosage of 150 mg/kg body weight.

Wound area analysis

A camera with a fixed height was used to take pictures to observe the wound-healing process of mice. A silicone ring was used as a scaffold to reduce the effect of rodent skin shrinkage on wound healing. The inner ring of the silicone ring (7 mm in diameter) was innovatively taken as a reference instead of a ruler to avoid stereological differences in observation.

Photographs of the wound were taken every other day until it healed. The inner circle of the silicone splint was taken as a reference. ImageJ (version 1.8) software (National Institutes of Health, USA) was used to track the wound edge, and the actual wound area was calculated. The relative wound area was calculated as the proportion of the actual wound area at a certain time to the original wound area (immediate postoperative wound area).

Histological and immunochemical analysis

Skin samples were collected on days 6 and 14 after surgery. Half of the samples were fixed in 4% paraformaldehyde, and the other half was frozen with liquid nitrogen.

After the fixed samples were gradually dehydrated, they were embedded in paraffin and cut into 4 µm slices perpendicular to the wound surface.

Hematoxylin and eosin (H&E) staining was conducted in accordance with established procedures. Tissue sections were dewaxed, hydrated, and stained with hematoxylin (G1004, Servicebio, China) for nuclear staining. After being rinsed, the sections were stained with eosin (E8090, Solarbio, China) for cytoplasmic and extracellular matrix staining. Subsequently, the sections were dehydrated, cleared, and mounted with a mounting medium.

Masson trichrome staining was performed using a Masson Trichrome Stain Kit (G1340, Solarbio, China) in accordance with the manufacturer’s instructions. Tissue sections were dewaxed, hydrated, and stained with Weigert’s iron hematoxylin for nuclear staining. After being rinsed, the sections were differentiated in acid alcohol and stained with Biebrich scarlet-acid fuchsin for collagen and muscle fiber staining. After being rinsed again, the sections were treated with phosphomolybdic acid to differentiate the fibers. They were then stained with aniline blue for cytoplasm. Finally, the sections were dehydrated, cleared, and mounted.

The stained sections were examined under a light microscope for histological analysis. The thickness of the epidermis and dermis was measured quantitatively by ImageJ (version) 1.8 software (National Institutes of Health, USA).

The expression levels of HIF-1α and PCNA were evaluated by immunochemistry. First, tissue sections were fixed, permeabilized, and blocked. Second, primary antibodies were applied, incubated overnight, washed, and then detected with labeled secondary antibodies. They were washed again, and signals were visualized with a 3,3N-diaminobenzidine tertrahydrochloride horseradish peroxidase color development kit (P0202, Beyotime, China). The dilution ratio of HIF-1α (20960-1-AP, Proteintech, USA) and PCNA (10205-2-AP, Proteintech, USA) was 1:100 and that of horseradish peroxidase (HRP)-labeled Goat Anti-Rabbit immunoglobulin G (H+L; A0208, Beyotime, China) was 1:50. After being mounted, the immunohistochemically stained sections were photographed using an optical microscope. ImageJ (version 1.8) software (National Institutes of Health, USA) was used to measure the mean optical density of HIF-1α and PCNA in new epidermis.

Western blot (WB)

The frozen sample was cut into pieces, grinded in liquid nitrogen, and lysed in radioimmunoprecipitation assay (RIPA) lysis buffer (R0010, Solarbio, Beijing, China) containing protease and phosphatase inhibitor cocktail (P1260, Solarbio, Beijing, China). The protein content was measured using a bicinchoninic acid protein assay kit (PC0020, Solarbio, Beijing, China). An equal amount of protein in each group was separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and transferred to polyvinylidene fluoride (PVDF) membranes (YA1701, Solarbio, Beijing, China). The membrane was blocked with 5% skim milk in tris-buffered saline with tween 20 and incubated with the primary antibody overnight. The source and dilution ratio of the antibody was as follows: Integrin β1 (1:1000, 26918-1-AP, Proteintech, Wuhan, China), K14 (1:1000, 22221-1-AP, Proteintech, Wuhan, China), K1 (1:800, 16848-1-AP, Proteintech, Wuhan, China), K10 (1:1000, 18343-1-AP, Proteintech, Wuhan, China), Notch 1 (Notch Intracellular Domain, NICD) (1:2000, 20687-1-AP, Proteintech, Wuhan, China), glyceraldehyde-3-phosphate dehydrogenase (GAPDH) (1:2000, 10494-1-AP, Proteintech, Wuhan, China), β-actin (1:5000, 20536-1-AP, Proteintech, Wuhan, China), VEGF (1:1000, 19003-1-AP, Proteintech, Wuhan, China), and CD31 (1:1000, 28083-1-AP, Proteintech, Wuhan, China). After being washed, the membranes were incubated with HRP-labeled secondary antibodies (1:1000, A0208, Beyotime, China). The membranes were washed again, and signals were visualized with BeyoECL Plus (P0018M, Beyotime, Shanghai, China). Two internal controls, GAPDH and β-actin, were chosen in this experiment. Images were acquired using a gel imaging system, and the obtained images were later analyzed for gray values using ImageJ (version 1.8) software (National Institutes of Health, USA).

Cell experiment grouping scheme and treatment measures

The cell line used in this study was HaCaT (Catalog number: iCell-h066, Supplier: Saibaikang, Shanghai, China). Short tandem repeat profiling was used to verify the identity of the cell line to ensure its authenticity and integrity. Mycoplasma assays were performed to confirm the absence of mycoplasma contamination in the cell lines before their use in the experiments.

HaCaT cells were divided into control (Glu 5.5 mM), high-glucose (Glu30 mM), and high-glucose + drug (Glu30 mM + FG-4592) groups, with a treatment concentration of FG-4592 set at 10 µM. Following 48 h of treatment period, the cells from each group were harvested, and protein was extracted.

Co-immunoprecipitation (Co-IP)

Co-IP was used to clarify the interaction between HIF-1a and NICD. As a positive control group, a certain amount of HaCaT cell protein was collected for WB. HIF-1a antibodies (1:200, sc-13515, Santa Cruz, USA) were pre-processed by mixing with magnetic beads (HY-K0204, MCE, China) for 2 h at 4°C. After the antibodies-and-magnetic beads complex was washed, then a certain amount of HaCaT cell protein was added to them for 2 h at 4°C. After the protein-antibodies-magnetic beads complex was washed, the immunoprecipitated proteins were recovered from the beads by boiling for 5 min in a sample buffer and then analyzed by immunoblotting as previously described.

Immunofluorescence (IF)

Cells were seeded onto coverslips and grown to an appropriate density. After being washed with PBS, the cells were fixed with 4% paraformaldehyde at room temperature for 15 min. They were then permeabilized with 0.1% Triton X-100 for 5 min, followed by washing with PBS. Afterward, the cells were blocked with 5% BSA for 30 min. The primary antibodies for HIF-1α (1:100, 66730-1-Ig, Proteintech, China) and Notch1 ICD (1:100, PAB35376, Bioswamp, China) were incubated overnight at 4°C. After the cells were washed with PBS, Alexa Fluor 594-conjugated Goat Anti-Mouse (1:200, SAB51359, Bioswamp, Wuhan, China) and HyperFluor 488-conjugated Goat anti-Rabbit (1:200, SAB43742, Bioswamp, Wuhan, China) were added, and the cells were incubated in the dark at room temperature for 1 h. After being washed with PBS, the cells were stained with DAPI (C0065, Solarbio, Beijing, China) and mounted. Fluorescent signals were observed using a fluorescence microscope (Leica DFC7000T, Germany).

Statistical analysis

Statistical comparisons and drawing were conducted in GraphPad Prism (version 6.0) software (GraphPad Software, San Diego, CA, USA). All data were first subjected to tests for normality and homogeneity of variances. Then, data that conformed to a normal distribution by the Shapiro-Wilk test and had homogeneous variances were presented as mean ± standard deviation. One-way analysis of variance (ANOVA) was used for the comparison between groups, and Fisher’s Least Significant Difference test was used for post-test and multiple comparisons of one-way ANOVA. All statistical tests were two-sided, with a P < 0.05 considered statistically significant. Fig Draw software (version 2.0, www.figdraw.com) was used for data visualization.

Roxadustat (FG-4592) accelerates the delayed wound healing of diabetic mice

The speed of wound closure is the most important parameter when evaluating the effect of roxadustat (FG-4592) on diabetic wound healing. As shown in Figures 1a and b, the relative wound area among all groups (P < 0.05) on days 4, 8, and 12 had significant differences.

Close

Figure 1:

Roxadustat (FG-4592) accelerates the delayed wound healing of diabetic mice. (a) Photos of wounds in each group. (b) Relative wound area of each group. (c) Blood glucose of each group. (d) Body weight of each group. (e) Weight change compared with the weight at the beginning of diabetes model induction in each group. n = 6 in each group. ^✶P < 0.05.

Export to PPT

First, the relative wound area of the diabetes group was significantly larger than that of the control group on days 4 (P < 0.05), 8 (P < 0.05), and 12 (P < 0.05) after injury. This finding indicated that the diabetic wound modeling was successful because this model clearly demonstrated the delayed healing of diabetic wounds. The relative wound area of the diabetes + FG-4592 group was significantly smaller than that of the diabetes group on days 4 (P < 0.05), 8 (P < 0.05), and 12 (P < 0.05) after injury, indicating that FG-4592 accelerated the delayed wound healing of diabetic mice.

Blood glucose was continuously monitored during the healing process. As shown in Figure 1c, the blood glucose level of the control group fluctuated between 5 and 11 mmol/L, whereas those of diabetes and diabetes + FG-4592 groups were stably higher than 16.7 mmol/L. No significant difference was found between the two groups, indicating that the diabetes model was stable and reliable and that FG-4592 had no effect on blood glucose.

The weight of the mice was continuously monitored before and after diabetes induction and during wound healing. As shown in Figure 1d, the body weight of the control group was higher than that of the diabetes and diabetes + FG-4592 groups (P < 0.05). The weight change in each group was compared with the weight at the beginning of diabetes induction, as shown in Figure 1e. The weight of the control group increased steadily. Conversely, the weight of the diabetes and diabetes + FG-4592 groups clearly decreased compared with that at the beginning of the diabetes induction. In the daily management of mice, the diabetic mice demonstrated obvious symptoms of polydipsia, polyphagia, and polyuria.

Roxadustat (FG-4592) improves the wound-healing quality of diabetic mice

H&E staining and Masson staining were applied to evaluate the healing quality. As shown in Figures 2a and b, in comparison with the control group, the newly healed skin of the diabetes group in the epidermis and dermis was thin, with few blood vessels, lessened collagen deposition, and reduced wound contraction. Meanwhile, the diabetes + FG-4592 group had an irregularly thickened epidermal layer, more collagen deposition, and more considerable wound contraction than the diabetes group. These findings are consistent with the phenomenon shown in Figure 1a. Quantitative analysis showed that the newly healed skin of normal mice (451.4 ± 47.80µm) was the thickest, followed by the diabetes + FG-4592 group (286.2 ± 79.99µm), whereas that of the diabetes group (151.9 ± 26.09 µm) was the thinnest [Figure 2c]. Significant differences were found in the thickness of newly healed skin among all groups. The skin of the diabetes group was significantly thinner than that of the control group (P < 0.05) and the diabetes + FG-4592 group (P < 0.05).

Close

Figure 2:

Roxadustat (FG-4592) improves the wound healing quality of diabetic mice. H&E staining (a) and Masson staining (b) of wound on day 14 after surgery. Bar = 500 μm. Images in the first and second rows are in the same group and both on day 14 after surgery, yellow arrows indicate epidermis, and white arrows indicate dermis. (c) Skin thickness of healed skin in each group at day 14 after surgery. n = 6 (3 from H&E staining sections and 3 from Masson staining sections). ^✶P < 0.05. H&E: Hematoxylin and eosin.

Export to PPT

Roxadustat (FG-4592) upregulates HIF-1 signaling and reverses the depressed proliferation of keratinocytes in diabetic mice

Promoting cell proliferation is the key to accelerate wound healing. Immunohistochemical staining was performed to confirm the effect of roxadustat (FG-4592) on HIF-1 signaling and cell proliferation. As shown in Figure 3a, in the diabetes + FG-4592 group, the HIF-1α highly-expressed region was mainly concentrated in the irregular thickening area of the epidermis. In the control and diabetes groups, the brown areas were relatively few and scattered in the epidermis. HIF-1α was expressed in dermal fibroblasts, but the difference was not as pronounced as in the epidermis.

Close

Figure 3:

Roxadustat (FG-4592) upregulates HIF-1 signaling and reverses the depressed proliferation of keratinocytes in diabetic mice. Immunohistochemical image of (a) HIF-1α and (b) PCNA on day 14 after surgery; bar = (a) 200 μm and (b) 500 μm for first row. The image in the second row is an enlarged version of the area in the black box in the first row; bar = 50 μm in the enlarged picture. Mean optical density of HIF-1α (c) and PCNA (d) in the epidermis layer. n = 3 wound samples in each group. ^✶P < 0.05. HIF-1: Hypoxia-inducible factor 1, PCNA: Proliferating cell nuclear antigen.

Export to PPT

As shown in Figure 3b, in the diabetes + FG-4592 group, the PCNA+ cells were densely distributed in the newly healed epidermis, almost covering all layers of the epidermis. In the diabetes group, the PCNA+ cells were the fewest and scattered in the already thin epidermis. In the control group, the epidermis thickness was uniform, and the PCNA+ cells were mainly distributed in the basal layer. The immunohistochemical images appeared to exhibit overdevelopment, often resulting in excessively dark staining. While PCNA is primarily expressed in the cell nucleus, a notably darkened staining was observed in the cytoplasm of the diabetes + FG-4592 group, and this finding may be attributed to edge effects due to overdevelopment.

As shown in Figures 3c and d, the expression of levels HIF-1α and PCNA significantly differed among groups. The HIF-1α expression in the diabetes group was significantly lower than that in the control group (P < 0.05) and diabetes + FG-4592 group (P < 0.05). Similarly, the PCNA expression in the diabetes group was significantly lower than that in the control group (P < 0.05) and diabetes + FG-4592 group (P < 0.05).

Roxadustat (FG-4592) promotes dedifferentiation of keratinocytes and angiogenesis in diabetic mice

The expression levels of integrin β1, K14, K10, K1, Notch1 NICD, CD31, and VEGF were evaluated by WB to investigate the effect of roxadustat (FG-4592) on keratinocyte differentiation and angiogenesis in diabetic wounds.

As shown in Figure 4a-4f, the expression levels of K14 and integrin β1, markers of basal cells, in the diabetes group were significantly reduced compared with those in the control group (P < 0.05). The expression of K10, the marker of differentiated cells, in the diabetes group significantly increased (P < 0.05), whereas that of K1 decreased (P < 0.05). The treatment of FG-4592 in diabetic mice induced keratinocytes to remain in an undifferentiated state (K14+ and integrin β1+) in diabetic wounds. The activation of Notch1 signaling was detected to understand the underlying mechanisms. The diabetic group demonstrated hyperactivation of Notch1 signaling compared with the control group (P < 0.05). Meanwhile, the FG-4592-treated group exhibited downregulation of Notch1 signaling compared with the diabetes group (P < 0.05). The trend was almost the same on days 6 and 14 post-injury. The change trends of K1 and K10 were not the same, which needs further investigation.

Close

Figure 4:

Roxadustat (FG-4592) promotes dedifferentiation of keratinocytes and angiogenesis in diabetic mice. (a) Expression levels of integrin β1, K14, K10, K1, and Notch1 NICD evaluated by Western blot in the middle and at the end of wound healing. (b-f) Corresponding quantitative analysis. (g) CD31 and VEGF expression levels evaluated by Western blot. (h and i) Corresponding quantitative analysis. n = 3 in each group. ^✶P < 0.05. NICD: Notch Intracellular Domain, VEGF: Vascular endothelial growth factor, K14: Keratin 14, K10: Keratin 10, K1: Keratin 1.

Export to PPT

As shown in Figure 4g-4i, the expression levels of VEGF and CD31 in the diabetes group notably decreased in comparison with those of the control group (P < 0.05). Compared with the diabetes group, the FG-4592-treated group showed a significant increase in the VEGF and CD31 expression levels (P < 0.05).

Roxadustat (FG-4592) promotes dedifferentiation in keratinocytes through interaction between HIF-1a and NICD

HaCaT cells were treated with 30 mM glucose for 48 h to simulate high glucose in vitro to further validate the dedifferentiation effect of FG-4592 on keratinocytes. The reversal effect of 10 µM FG-4592 on high-glucose-induced Notch signaling overactivation and keratinocyte over-differentiation was detected. As shown in Figures 5a and b, IF and WB showed that compared with the control group, the high-glucose group had decreased expression levels of HIF-1a, integrin β1, and K14 and an increased expression of Notch1 NICD. Compared with the high-glucose group, the high-glucose + drug group showed an increase in the expression levels of HIF-1a, integrin β1, and K14 and a decrease in the expression of Notch1 NICD. The differences were statistically significant according to quantitative analysis through WB [Figure 5c-f].

Close

Figure 5:

Roxadustat (FG-4592) promotes dedifferentiation through interaction between HIF-1α and NICD. (a) Immunofluorescence showing the expression and co-localization of HIF-1α and NICD in HaCaT cells under different conditions; bar = 100 μm. (b) Western blot showing the expression levels of HIF-1α, NICD, K14, and integrin-β1 in HaCaT cells under different conditions. (c-f) Quantitative analysis of WB. n = 3 in each group; ^✶P < 0.05. (g) Co-IP confirming the interaction between HIF-1α and NICD. HIF-1: Hypoxia-inducible factor 1, NICD: Notch intracellular domain, WB: Western blot, Co-IP: Co-immunoprecipitation, K14: Keratin 14.

Export to PPT

Co-localization of HIF-1a and Notch1 NICD [Figure 5a] was observed in IF staining and co-immunoprecipitation further confirmed the interaction between HIF-1a and Notch1 NICD [Figure 5g].